EP0829048A2

EP0829048A2 - Programmable instruction trap system and method

Info

Publication number: EP0829048A2
Application number: EP96916908A
Authority: EP
Inventors: Sunil Savkar; Gene Shen; Farnad Sajjadian; Mike Shebanow
Original assignee: HAL Computer Systems Inc
Current assignee: Fujitsu Ltd
Priority date: 1995-06-01
Filing date: 1996-05-31
Publication date: 1998-03-18
Anticipated expiration: 2016-05-31
Also published as: DE69626282T2; JP3604697B2; WO1996038789A3; JP2000504444A; DE69626282D1; EP0829048B1; US5896526A; WO1996038789A2

Abstract

A system and method providing a programmable hardware device within a CPU. The programmable hardware device permits a plurality of instructions to be trapped before they are executed. The instructions that are to be trapped are programmable to provide flexibility during CPU debugging and to ensure that a variety of application programs can be properly executed by the CPU. The system must also provide a means for permitting a trapped instruction to be emulated and/or to be executed serially.

Description

PROGRAMMABLE INSTRUCTION TRAP SYSTEM AND METHOD

BACKGROUND OF THE INVENTION

Related Applications

The subject matter of this application is related to the following applications: application Serial Number entitled "METHOD

AND APPARATUS FOR ROTATING ACTIVE INSTRUCTIONS IN A PARALLEL DATA PROCESSOR" filed on June 1, 1995 by Sunil Savkar, Michael C. Shebanow, Gene W. Shen, and Farnad Sajjadian; application Serial Number 08/388,602 entitled "INSTRUCTION FLOW CONTROL CIRCUIT FOR SUPERSCALER MICROPROCESSOR" filed on February 14, 1995 by Takeshi Kitahara; application Serial Number 08/388,389 entitled "ADDRESSING

METHOD FOR EXECUTING LOAD INSTRUCTIONS OUT OF ORDER WITH RESPECT TO STORE INSTRUCTIONS" filed on February 14, 1995 by Michael A. Simone and Michael C. Shebanow; application Serial Number 08/388,606 entitled "METHOD AND APPARATUS FOR EFFICIENTLY WRITING RESULTS TO RENAMED REGISTERS" filed on February 14, 1995 by DeForest W. Tovey, Michael C. Shebanow and John Gmuender; application Serial Number 08/388,364 entitled "METHOD AND APPARATUS FOR COORDINATING THE USE OF PHYSICAL REGISTERS IN A MICROPROCESSOR" filed on February 14, 1995 by DeForest W. Tovey, Michael C. Shebanow and John Gmuender; application Serial Number entitled "PROCESSOR

STRUCTURE AND METHOD FOR TRACKING INSTRUCTION STATUS TO MAINTAIN PRECISE STATE" filed on February 14, 1995 by Gene W. Shen, John Szeto, Niteen A. Patkar and Michael C. Shebanow; application Serial Number entitled "PARALLEL

ACCESS MICRO-TLB TO SPEED UP ADDRESS TRANSLATION" filed on March 3, 1995 by Chih-Wei David Chang, Kioumars Dawallu, Joel F. Boney, Ming-Ying Li and Jen-Hong Charles Chen; application Serial Number entitled "LOOKASIDE

BUFFER FOR ADDRESS TRANSLATION IN A COMPUTER SYSTEM" filed on March 3, 1995 by Leon Kuo-Liang Peng, Yolin Lih and Chih-Wei David Chang; application Serial Number 08/397,893 entitled "RECLAMATION OF PROCESSOR RESOURCES IN A DATA PROCESSOR" filed on March 3, 1995 by Michael C. Shebanow, Gene W. Shen, Ravi Swami, Niteen Patkar; application Serial Number 08/397,891 entitled "METHOD AND APPARATUS FOR SELECTING INSTRUCTIONS FROM ONES READY TO EXECUTE" filed on March 3, 1995 by Michael C. Shebanow, John Gmuender, Michael A. Si one, John R.F.S. Szeto, Takumi Maruyama and DeForest W. Tovey; application Serial Number 08/397,911 entitled "HARDWARE SUPPORT FOR FAST SOFTWARE EMULATION OF UNIMPLEMENTED

INSTRUCTIONS" filed on March 3, 1995 by Shalesh Thusoo, Farnad Sajjadian, Jaspal Kohli, and Niteen Patkar; application Serial Number 08/398,284 entitled "METHOD AND APPARATUS FOR ACCELERATING CONTROL TRANSFER RETURNS" filed on March 3, 1995 by Akiro Katsuno, Sunil Savkar and Michael C. Shebanow; application Serial Number 08/398,066 entitled "METHODS FOR UPDATING FETCH PROGRAM COUNTER" filed on March 3, 1995 by Akira Katsuno, Niteen A. Patkar, Sunil Savkar and Michael C. Shebanow; application Serial Number 08/397,910 entitled "METHOD AND APPARATUS FOR PRIORITIZING AND HANDLING ERRORS IN A COMPUTER SYSTEM" filed on March 3, 1995 by Chih-Wei David Chang, Joel Fredrick Boney and Jaspal Kohli; application Serial Number 08/398,151 entitled "METHOD AND APPARATUS FOR RAPID EXECUTION OF CONTROL TRANSFER INSTRUCTIONS" filed on March 3, 1995 by Sunil W. Savkar; application Serial Number 08/397,800 entitled "METHOD AND APPARATUS FOR GENERATING A ZERO BIT STATUS FLAG IN A MICROPROCESSOR" filed on March 3, 1995 by Michael Simone; application Serial Number 08/397,912 entitled "ECC

PROTECTED MEMORY ORGANIZATION WITH PIPELINED READ-MODIFY-WRITE ACCESS" filed on March 3, 1995 by Chien Chen and Yizhi Lu; and application Serial Number 08/398,299 entitled "PROCESSOR STRUCTURE AND METHOD FOR TRACKING INSTRUCTION STATUS TO MAINTAIN PRECISE STATE" filed on March 3, 1995 by Chien Chen, John R.F.S, Szeto, Niteen A. Patkar, Michael C. Shebanow, Hideki Osone, Taku i Maruyama and Michael A. Simone; each of the above applications are incorporated herein by reference in their entirety.

1. Field of the Invention

This invention relates generally to data processors. More particularly, this invention relates to a system and method for providing programmable hardware instruction traps in a microprocessor.

2. Description of Background Art

When designing a data processor it is important to rigorously test the design before tape-out. Generally, such testing attempts to identify errors in functional characteristics of the data processor. However, due to time limitations it is not always feasible to fully test all functional characteristics of the data processor design before tape-out. Tape-out occurs when the physical database design for manufacturing the parts, e.g., the data processor, is released. Occasionally, functional errors in the design of the data processor are not discovered until after tape-out. In this circumstance it is expensive to correct the design and produce new data processor parts . Functional errors can also occur because of a flaw in manufacturing that effects certain instructions .

A related limitation occurs with superscalar central processing units (CPUs) . Many software application programs are not developed for use in a superscalar environment. When a superscalar CPU executes instructions in such an application program some instructions may not perform in the manner they were designed.

What is needed is a system and method for providing a programmable hardware device within a CPU that permits a plurality of instructions to be trapped before they are executed. The instructions that are to be trapped are^' programmable to provide flexibility during CPU debugging and to ensure that a variety of application programs can be properly executed by the CPU. The system must also provide a means for permitting a trapped instruction to be emulated and/or to be executed serially.

SUMMARY OF THE INVENTION

The invention is a system and method that provides a programmable hardware device within a CPU. The programmable hardware device permits a plurality of instructions to be trapped before they are executed. The instructions that are to be trapped are programmable to provide flexibility during CPU debugging and to ensure that a variety of application programs can be properly executed by the CPU. The system must also provide a means for permitting a trapped instruction to be emulated and/or to be executed serially.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is an illustration of a computer environment in which a preferred embodiment of the present invention resides.

Figure 2 is a more detailed illustration of a superscalar central processing unit of the preferred embodiment of the present invention.

Figure 3 is a more detailed illustration of a programmable instruction trap unit according to the preferred embodiment of the present invention.

Figure 4 is a more detailed illustration of a programmable instruction trap unit according to the preferred embodiment of the present invention. Figure 5 is a more detailed illustration of an Instruction Trap Register Value (ITRV) and an Instruction Trap Register Mask (ITRM) according to the preferred embodiment of the present invention.

Figure 6 is a more detailed illustration of a bit register according to the preferred embodiment of the present invention.

Figure 7 is a flow diagram describing the method of trapping an instruction according to the preferred embodiment of the present invention.

Figure 8 is a flow diagram describing the method of testing an instruction according to the preferred embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

A preferred embodiment of the present invention is now described with reference to the figures where like reference numbers indicate identical or functionally similar elements. Also in the figures, the left most digit of each reference number corresponds to the figure in which the reference number is first used.

Figure 1 is an illustration of a computer environment in which a preferred embodiment of the present invention resides . A conventional memory unit 104 and conventional input/output

(I/O) devices 106 communicate with a data processor 102. In the preferred embodiment the processor is a HaL Rl processor developed by HaL Computer Systems, Campbell, California. The HaL Rl processor implements the SPARC V9 64 bit instruction set architecture developed by SPARC International, Inc., Menlo Park, California. The processor 102 includes a superscalar CPU 108, a memory management unit (MMU) , and four cache memory units 110, 112.

The cache memory units 110, 112 preferably include two 64 kilobyte (KB) conventional data caches and two 64 KB conventional instruction caches. The data caches 110 provide data to the superscalar CPU 108 and the MMU 114. Both sets of cache memory units 110, 112 are virtually indexed and tagged. Each cache chip 110, 112 contains 64 KB of data storage, organized as four sets. Each 64 KB cache chip 110, 112, can service two independent requests from the superscalar CPU 108. The interface between the superscalar CPU 108 and the caches 110, 112 is non-blocking in that the superscalar CPU 108 can access the cache 110, 112, at the same time a cache line is being refilled or flushed. Additional detail concerning the cache memory units 110, 112 is set forth in Chen, Chien, Y. Lu, A. Wond; The Microarchi tecture of the HaL Cache, Compcon Proceedings 1995, which is incorporated herein by reference in its entirety.

The function of the MMU 114 includes controlling memory management and data coherence, interfacing to memory 104 and I/O devices 106, and error handling, for the processor 102. The MMU 114 includes three levels of address spaces: (1) a virtual address space for the processor 102, (2) a logical address space for I/O devices 106 and a diagnostic processor, and (3) a physical address space for memory 104. These hierarchical address spaces provide an efficient mechanism for managing the large address space. A more detailed description of the MMU 114 used in the preferred embodiment is described in David Chang, D. Lyon, C. Chen, L. Peng, M. Massoumi, M. Hakimi, S. Iyengar, E. Li, R. Remedios, Microarchi tecture of HaL ' s Memory Management Uni t, Compcon Proceedings 1995, which is incorporated herein by reference in its entirety. Figure 2 is a more detailed illustration of a superscalar CPU 108 of the preferred embodiment of the present invention. The processor 102 implements a four-stage pipeline for fixed- point instructions and a six-stage pipeline for "LOAD" instructions. The superscalar CPU 108 includes an instruction fetch unit 202, a programmable instruction trap unit 204, an instruction issue unit 206, an instruction execution unit 218, a precise state unit 220, a branch history table 208, a return prediction table 210, a prefetch buffer 212, a recoded instruction cache 214 and a branch unit 216. The fetch unit 202 requests and receives four instructions from either the off-chip 64 KB instruction cache 112, shown as data lines 201 in figure 2, a prefetch buffer 212, or a recoded instruction cache 214. Four instructions are then sent to a programmable instruction trap unit 204 to determine if any of the instructions should be trapped. The programmable instruction trap unit 204 is described in greater detail below with reference to figures 3-8. The branch history table 208 contains a plurality of counters that are used to predict the direction of execution branches. The return prediction table 210 is used to predict the return address of executed instructions. The branch unit 216 uses information from the issue unit 206, the branch history table 208, and the return prediction table 210, to determine if a control transfer instruction is within the current issue window, and further modifies the program counter if the control transfer instruction is taken. The issue unit 206 determines the available computer resources and issue constraints, e.g., should an instruction be synchronized. All instructions are issued by the issue unit 206 in the order they would have issued in a traditional serial processor and are sent to the instruction execution unit 218. The instruction execution unit 218 executes the instructions, although not necessarily in the order in which they were issued. The instruction execution unit 218 outputs the results of the instruction execution on data bus 222. A more detailed description of the superscalar CPU 108 of the preferred embodiment is described in Niteen Patkar, A. Katsuno, S. Li, T. Maruyama, S. Savkar, M. Simone, G. Shen, R. Swami, D. Tovey, Microarchi tecture of the HaL PMl CPU, Compcon Proceedings 1995. The present invention lies in the programmable instruction trap unit 204 which is herein described.

Figure 3 is a more detailed illustration of a programmable instruction trap unit 204 according to the preferred embodiment of the present invention. The programmable instruction trap unit 204 preferably receives four instructions from the fetch unit 202. The instructions are shown on figure 3 as INST00, INST01, INST10, INSTIL Each instruction is sent to an instruction trap logic unit (ITLU) 302. The instruction trap logic unit also receives signals from preferably four instruction trap register's (ITRV) 304, four instruction trap register masks (ITRM) 306, and four control registers (CR) 308. A more detailed description of the instruction trap logic unit 302, ITRM 304, ITRM 306 and CR 308 are set forth below with respect to figures 4-8. Figure 4 is a more detailed illustration of a programmable instruction trap unit according to the preferred embodiment of the present invention. The instruction trap logic unit 302 preferably includes four Exclusive NOR logic devices 404A-D, where one Exclusive NOR is associated with each ITRV 304. Each Exclusive NOR device 404 compares each inverted bit of the instruction with a corresponding bit from the associated ITRV, e.g., ITRVO 410, and for each bit the NOR device 404 outputs a value indicating whether a match has occurred. It is envisioned that other techniques can be used to compare each bit of the instruction, INST00, with each bit in an ITRV 304. The method for performing this comparison is described in greater detail below with respect to figures 7-8.

The instruction trap logic unit 302 preferably includes four devices 406A-D that compare each bit output by an associated Exclusive NOR device 404 with a bit value in an ITRM 306. Each ITRM 306 contains a plurality of bits containing a "care" or "don't care" value. Each of the four devices 406A-D compares a bit that is output by its associated Exclusive NOR device and a bit in its associated ITRM 306. Each device 406 outputs a one bit value representing whether all of the instruction bits match the associated ITRV-ITRM register pair. Preferably, each device 406 is an A032x2 device. This device has two sets of 32 bit inputs, one set coupled to each bit line of the associated Exclusive NOR 404 output, the other set coupled to the associated ITRM 306. The A032x2 includes 32 logical AND gates whose outputs are input into a logical OR gate. The method for performing this comparison is described in greater detail below with respect to figures 7-8. The CR's 308 are associated with a single ITRV-ITRM pair. For example, CR0 is associated with ITRVO and ITRM0. Each CR contains data representing the action that is to be taken if the instruction matches its associated ITRV-ITRM register pair. Although each control register can contain any number of bits, in the preferred embodiment each CR 308 contains two bits. One bit represents a synchronization action, the other bit represent an instruction trap. That is, if the bits in a control register are "00" and if the instruction matches the associated ITRV-ITRM register pair 304, 306, then no special action is taken. If the bits in a control register are "01" and if the instruction matches the associated ITRV-ITRM register pair 304, 306, then the programmable• instruction trap unit 204 will output a signal that causes the matched instruction to execute synchronously, i.e., all instructions that are issued before the matched instruction will have completed execution and the machine will be in a committed, i.e., known, state before the matched instruction executes. In addition, a synchronization instruction is issued, executed, completed and committed before any subsequent instructions are issued. If the bits in a control register are "10" and if the instruction matches the associated ITRV-ITRM register pair 304, 306, then the programmable instruction trap unit 204 will generate a non- synchronizing instruction issue trap, i.e., the instruction will take on the semantics of an instruction trap. If the bits in a control register are "11" and if the instruction matches the associated ITRV-ITRM register pair 304, 306, then the programmable instruction trap unit 204 will generate a synchronizing instruction issue trap, i.e., the instruction will be trapped and executed synchronously.

The instruction trap logic unit 302 includes a device for generating the values stored in the CR 308 associated with an ITRV-ITRM register pair 304, 306 that matches the received instruction, e.g., INST00. A variety of techniques can be utilized to generate this value. Preferably, two sets of four logical OR gates (one associated with each CR 308) receive an input from the CR 308. Each OR gate of the first set of OR gates receives two inputs: (1) the first bit of a CR 308, e.g., CR0, and (2) the one bit output from the A032x2 associated with the CR 308. Similarly, each OR gate of the second set of OR gates receives two inputs: (1) the second bit of a CR 308, e.g., CR0, and (2) the one bit output from the A032x2 associated with the CR 308. The output of all of the OR gates for each set of OR gates is supplied to a logical NAND gate, 408A-B. The method for generating the outputs of the instruction trap logic unit

302 is described in greater detail below with respect to figures 7-8.

The values in the ITRV's 304, the ITRM's 306, and the CR's 308 are programmable. These values can be programmed by a user to trap instructions and cause certain instructions to be executed synchronously, as described above. The program ability feature of the present invention provides an efficient technique for trapping and synchronizing instruction in order to correct functional errors in processor instructions and ensure the proper operation of application programs by executing certain instructions synchronously, i.e., those instructions stored in the ITRV 304, ITRM 306. A variety of techniques can be used to program the ITRV's 304, the ITRM's 306, and the CR's 308 (collectively called the programmable registers) . For example, the programmable registers can be coupled to a data bus and to control logic that permits the programmable registers to receive data from the data bus. In the preferred embodiment the data is "scanned-in" during the computer boot-up operation. The programmable registers are programmed using a technique similar to shift registers.

Figure 5 is a more detailed illustration of an ITRV 304 and ITRM 306 according to the preferred embodiment of the present invention. Each programmable register 410 contains a plurality of bit registers 502. Preferably, each ITRV 304 and each ITRM 306 include 27 bit registers 502 while each CR 306 includes 2 bit registers 502. Figure 5 is a representative ITRV 304 or an ITRM 306. The CR's 308 operate in the same manner with the exception that each CR 308 preferably includes two bit registers 502 as opposed to 27 bit registers 502, as discussed above. Each programmable register 410 receives a scan clock (SC) signal, a scan-enable (SE) signal, and scan-input (SI) signal, and generates a scan-output (SO) signal and a Q output signal. Figure 6 is a more detailed illustration of a bit register 502 according to the preferred embodiment of the present invention. A signal containing a bit value is received on the SI line. When the SC and SE signals are high, i.e., the signal represents a binary "1" value, the logical AND gate 602 generates a high signal. This signal is coupled to the gate of transistor 604. As a result, transistor 604 conducts thereby enabling the signal on the SI line to pass through the transistor 604 and have its logic value be maintained by two inverters 606. When the SC goes low while the SE signal remains high transistor 604 turns off and the logical AND gate 608 generates a high signal. The AND gate 608 is coupled to the gate of transistor 610. As a result, transistor 612 conducts and the inverted SI signal is again inverted by one of the inverters 610 and is output on the SO line. The signal is also output on the Q line that is coupled to either device 404A if the bit register is part of an ITRV 304 , an inverter and device 406A if the bit register is part of an ITRM 306, or an inverter and an OR gate if the bit register is part of a CR 308. The SO line is coupled to the next bit register SI line. The SO line of the last bit register of ITRV 304 is coupled to the first bit register of the first ITRM 306. The last bit register of ITRM 306 is coupled to the first bit register of the first CR 308. During the computer initialization process data representing the programmable register values are systematically input into the first SI line 412 and shifted through all of the bit registers until each programmable register contains the desired values, i.e., the bit register values are scanned-in.

The method of the present invention will now be described. Figure 7 is a flow diagram describing the method of trapping an instruction according to the preferred embodiment of the present invention. Initially, the ITRV's 304, the ITRM's 306, and the CR's 308 are programmed. As described above, these programmable registers can be programmed individually 704, 706, 708, using a data bus and control signals. Preferably, the programmable registers are programmed during computer initialization using the scan-in, i.e., bit shifting, technique described above with reference to figures 5-6. After the programmable registers are programmed 702 instructions are tested 710. The method for testing instructions is set forth below with respect to figure 8.

Figure 8 is a flow diagram describing the method of testing an instruction according to the preferred embodiment of the present invention. Preferably four instructions are received 802 by the programmable instruction trap unit 204 from the fetch unit 202, as illustrated in figure 2. Each instruction is sent to a separate instruction trap logic unit 302. For example, instruction INST00 is sent to instruction trap logic unit 302A. The instruction trap logic unit 302A inverts each bit in the instruction using inverter 402. The inverted instruction bits are received by four exclusive NOR device 404A-D, described above with reference to figure 4. Each exclusive NOR device 404 compares 804 each inverted instruction bit with a corresponding bit in one of the ITRV's 304 that was previously programmed. For example, the exclusive NOR device 404A, compares instruction bit 0 with bit register 0 of ITRVO, as illustrated in figures 4- 6. If the inverted instruction bit matches the associated bit in one of the ITRV's 304 then the exclusive NOR generates a signal representing a logical "1". Otherwise, the exclusive NOR generates a signal representing a logical "0". Since the exclusive NOR device is comparing the associated ITRV bit register 502 and an inverted bit of the instruction, if the exclusive NOR generates a signal representing a logical "0" then the ITRV bit matches the instruction bit. Each inverted bit of the instruction, INST00, is compared to each associated bit of each ITRV 304. Preferably, the result of these comparisons is four 27-bit signals representing the result of the exclusive NOR comparison of devices 404A-D.

Preferably, each bit in each set of 27-bit signals is compared 806 with a bit register 502 in an associated ITRM 306 using an A032x2 device 406A-D. Each A032x2 device 406A-D consists of at least 27 logical AND gates whose outputs are input into a logical OR gate. For example, the least significant bit, or the zero bit, that is output from device 404A is input into an AND gate along with the inverted zero bit register of ITRM0. The ITRM's 306 are a mask. Sometimes it is desirable to compare each instruction with a set of related instruction patterns. For example, it may be desirable to match all instructions having a certain combination of bits 5-9. In this situation the remaining bits, e.g., bits 0-4 and 10-26, should be masked because their value is irrelevant for this particular match. This matching is accomplished by placing a "don't care" value in the proper ITRM bit register. In the present invention, the "don't care" value is a logical "1". As illustrated in figure 4, the ITRM values are inverted before they are received by the A032x2. Accordingly, if a bit from device 404A is either zero, because the instruction bit matched the associated ITRV bit register, the output of the associated AND gate in the A032x2 406A will be zero. Similarly, if a bit from device 404A is a logical one, then the output of the AND will be zero only if the associated bit register in ITRM0 is masked, i.e., it's inverted value is equal to zero.

As discussed above, all of the AND gate outputs are coupled to an OR gate. If any AND gate output is a logical one then the output of the OR gate is a logical one. The output of the OR gate is the output of the A032x2 406. Therefore, the A032x2 406 generates an instruction match value (IMV) signal that represents a logical zero only if all of the bits from the instruction INSTOO either match the associated ITRV bits or are masked, i.e., if the instruction INSTOO matches the ITRV-ITRM 304, 306 register pair.

The programmable instruction trap unit 204 then determines 810 if any matches have occurred and if so it generates 812 the value in the associated CR 308. The process for achieving this is now described. The output from each A032x2 device 406 is received by two logical OR gates, e.g., OR gates 418, 420, as illustrated in figure 4. The second input to the first OR gate 418 is the value of the first bit register of the associated CR 308. The second input to the second OR gate 420 is the value of the second bit register of the associated CR 308. For example, if instruction INSTOO only matches the register pair ITRV0-ITRM0 then A032x2 406A generates a signal representing a logical zero at its output and the remaining A032x2 406B-D generate a signal representing a logical one at its output. The three OR gates associated with A032x2's 406B-D receive a binary one as one of its inputs. Accordingly, these three OR gates will output a signal representing a binary one. NAND gate 408A will generate a logical zero output only if the output of OR gate 418 generates a logical one. Since the output of device 406A is a logical zero, the output of OR gate 418 is a logical one only if the first bit register of CR0 contains a logical zero. Each set of four OR gates and one NAND gate 408 result in either (1) the value of the CR 308 associated with a matching register pair to be output by the NAND gates 408A-B, or (2) if the instruction INSTOO does not match any register pair, the NAND gates 408A-B output a "00" value representing that no special action is to be taken, as described above.

All four received instructions, i.e., INSTOO, INST01, INST10, and INSTll are similarly compared to each of the four ITRV registers and the ITRM registers.

It is possible for more than one register pair to match a single instruction. For example, both the register pair ITRV1- ITRM1 and the register pair ITRV2-ITRM2 can match a particular instruction, e.g., INSTOO. In such a circumstance a bit value of "1" in a CR bit register for one CR, e.g., CR2, will result in the associated NAND gate generating a signal representing a logical one. After all four instructions are concurrently tested the method continues 814. If more instructions are received 712 the process repeats. After all instructions have been tested new values for the ITRV's 304, ITRM's 306, and CR's 308 can be programmed 714 in using a data bus or preferably by rebooting the computer and scanning in new programmable register values, as described above, and then the process 702-714 repeats. While the invention has been particularly shown and described with reference to a preferred embodiment and several alternate embodiments thereof, it will be understood by persons skilled in the relevant art that various change in form and details can be made therein without departing from the spirit and scope of the invention.

Claims

CLAIMSWhat is claimed is:

1. A computer processing system comprising: an instruction fetch unit, disposed to receive a plurality of computer instructions, for retrieving said plurality of computer instructions; and a programmable instruction trap unit, coupled to said fetch unit, for comparing each of said computer instructions with one of a plurality of programmable values, and for generating an instruction trap control signal for each of said instructions representing a function to be performed for each of said computer instructions in response to said comparisons.

2. The computer processing system of claim 1, wherein said programmable instruction trap unit includes: a plurality of first programmable registers, each of said first programmable registers for storing a first value representing one or more trap instructions that a user wants to trap, said first value is programmable by said user; a plurality of first comparators, each disposed to receive one of said computer instructions and one of said programmable register values, for determining if said computer instruction matches said programmable register value; a plurality of programmable control registers, each of said programmable control registers associated with one of said first programmable registers, for storing a second value representing a function to be performed if said first comparator determines said computer instruction matches said programmable register value; and a control unit, coupled to said programmable control register, for generating a plurality of signals representing said second value if one of said first programmable registers matches one of said computer instructions.

3. The computer processing system of claim 2, wherein each of said first comparators comprises: a second comparator, disposed to receive said computer instruction; and a third comparator, coupled to said control unit and disposed to receive an output of said second comparator^".

4. The computer processing system of claim 3, wherein each of said first programmable registers includes : a value register, coupled to said second comparator, for storing data representing one or more of said trap instructions; a mask register coupled to said third comparator, for storing mask information associated with said value register, each bit of said mask information represents one of a mask bit indicator and a don't mask bit indicator.

5. The computer processing system of claim 4, wherein said second comparator compares each bit of said computer instruction with an associated bit of said value register and generates a first bit value representing the second comparator result.

6. The computer processing system of claim 5, wherein said third comparator compares each first bit value output with an associated bit from said mask register and generates a second bit value representing the third comparator result.

7. The computer processing system of claim 5, wherein said third comparator compares all of said first bit values output by said second comparator with all associated bits of said mask register and generates matching signal if all bit- comparisons match.

8. The computer processing system of claim 2, wherein four instructions are simultaneously compared with all of said first programmable registers.

9. A method for identifying instructions comprising the steps of: programming all bits in a plurality of first registers, a plurality of second registers, and a plurality of control registers; receiving a plurality of instructions having a plurality of bits; performing a first comparison of each bit of each instruction with an associated bit of each of said first registers; generating a first bit value for each of said first comparisons; performing a second comparison of each of said first bit values with an associated bit of said second register that is associated with said first bit value; and generating a signal representing the contents of an associated control register if said instruction satisfies said second comparison.